An eLLMentary RAG Evaluation Framework: How to Effectively Use Models as Judges